Effects of the New Basel Capital Accord on BankCapital Requirements for SMEs

EDWARD I. ALTMAN*

New York University

GABRIELE SABATO

University of Rome ‘‘La Sapienza’’

Abstract

Using data from three countries (US, Italy and Australia) and surveying related studies from several other

countries in Europe, we investigate the effects of the New Basel Capital Accord on bank capital requirements

for small and medium sized enterprises (SMEs). We find that, for all the countries, banks will have significant

benefits, in terms of lower capital requirements, when considering small and medium sized firms as retail

customers. But they will be obliged to use the Advanced IRB approach and to manage them on a pooled basis.

For SMEs as corporates, however, capital requirements will be slightly greater than under the existing Basel I

Capital Accord. We believe that most eligible banks will use a blended approach (considering some SMEs as

retail and some as corporate). Through a breakeven analysis, we find that for all of our countries, banking

organizations will be obliged to classify as retail at least 20% of their SME portfolio in order to maintain the

current capital requirement (8%).

Key words: SME finance, Basel II, bank capital requirements, retail bankingR

JEL classification: G21, G28

1. Introduction

Small and medium sized enterprises (SMEs) play a fundamental role in the economy of

many countries all over the world. For OECD members, the percentage of SMEs out of

the total number of firms is greater than 97%. They are a continuing source of dynamism

for the economy, producing three-fourths of the total jobs and often more than one-third

of the country’s GDP. Thanks to the simple structure of SMEs, they can respond quickly

to changing economic conditions and meet local customers’ needs, growing sometimes

into large and powerful corporations or failing within a short time of the firm’s inception.

Many public and private financial institutions, such as the World Bank1 or Governments2

* Please address correspondence to Edward I. Altman, NYU Salomon center, Leonard N. Stern School of

Business, New York University, 44 West 4th Street, New York, NY 10012, USA. E-mail: ealtman@

stern.nyu.edu.

1 BThe Challenge,^ World Bank Review of Small Business Activities (2001).

2 See, for example, US Small Business Administration, http://www.sba.gov, or EU policies and activities for

SMEs, http://europa.eu.int/comm/enterprise/entrepreneurship/promoting_entrepreneurship/index.htm.

Journal of Financial Services Research 28:1/2/3 15–42, 2005# 2005 Springer Science + Business Media, Inc. Manufactured in The Netherlands.

themselves, launch each year plans in order to sustain this essential player of nations’

economy. Borrowings, however, especially from commercial banks,3 remain undoubt-

edly the most important source of external SME financing.

Concerns have been raised that the new Basel Capital Accord (Basel II) will change

the way banks analyze credits, introducing new credit risk management techniques and

possibly reducing the lending activity toward SMEs. This is due to banks’ potential

perception that SMEs carry higher risk and, hence, higher capital requirements than

under Basel I. Many SME associations in different countries have publicly complained

about the new rules and many governments are now concerned. To reduce these

concerns, the European Commission published a report4 in which the SMEs access to

finance, for non-start-up companies, was analyzed. The report concluded that only 13%

of European SMEs consider access to finance as a major barrier. Approximately 80% of

them had at least one established credit line. However, this study’s conclusions did not

eliminate SME’s concerns about their future under the new Basel Accord.

From the beginning (1999) of the capital adequacy reform process, the Basel

Committee paid particular attention to the SME segment, mainly by having changed

the formulas to calculate risk weights linked to SMEs5 three times. In the last version

(June 2004, par. 232 and 273), banks are able to consider small and medium sized

enterprises as retail or as corporate entities, primarily based on their total exposure to this

entity. If total exposure is under "1 million ($1 million in some countries), SMEs can be

classified as retail, but another important qualitative requirement must be followed: the

credit must be managed as a retail exposure. This means that Bthe exposure must be one

of a large pool of exposures which are managed by the bank on a pooled basis.^ On the

other hand, if SME credits are classified as corporate, a special Bdiscount^ in the asset

correlation calculation is ensured for exposures to firms with under "50 million in sales.

Both approaches specify Bhaircuts^ which are based on the assumption that smaller firms

credit risks are less correlated as to default risk than larger corporates and less sensitive

to the business cycle. We also expect that the benefits shown for SMEs have been

motivated by nations’ concerns about lessening credit possibilities and the importance

that SME owners and employees have on voting results.

In this paper, we use data from three different countries (US, Italy and Australia),

considering the SME structure of each economy, in order to quantify the expected effect on

the bank capital requirements when considering a small firm as either retail or corporate.

In section 2, we review the extant European and US research literature on the

expected effects of Basel II on SMEs. This literature, all based on earlier versions of the

new Basel Accord, provides evidence of the potential benefits of the Basel II

implementation on the bank capital requirements, but it does not analyze effects on

SME lending by banking organizations. In section 3, we derive bank capital requirements

using formulas from the Advanced Internal Rating Based (A-IRB) approach as contained

3 See Berger and Udell (1998), Udell (2004) for a detailed discussion of lending to SMEs.

4 See European Commission, Enterprise Publications, B2003 observatory of European SMEs,^ for further

analyses.

5 Basel Committee on Banking Supervision, January 2001, April 2003, June 2004.

16 ALTMAN AND SABATO

in the last version of the Accord (June 2004). Realistic assumptions about maturity, loss

given default (LGD) and exposures are made. We are confident that alternative possible

realistic assumptions would not affect the results of the analysis. We also provide a

sensitivity analysis to show the effects on capital requirements of using different values

of LGD and maturity. For two of the countries (US and Australia), where data is

available, we also calculate the capital requirements for the large corporate segment

(where firms’ sales are more than " 50 million, $50 million in some countries), even if it

is not the focus of this paper.

In order to perform our analysis, we develop three different models to assign

probability of default (PD) estimates to the firms in our samples. In this way, we

specifically consider the differences between the structures of SMEs and their credit risk

attributes in different countries. Lastly, we summarize our results in terms of banks’

capital requirements for the three countries and we derive our main findings focusing on

the possible changes in the relationship between banks and SMEs and the expected

advantages and disadvantages of these changes. In section 4, we submit our conclusions.

2. Review of the relevant research literature

In this section, we review some of the most important works about the possible effects of

the Basel II implementation on small and medium sized enterprises. The literature

generally finds that for both Europe and the US, the new Basel Capital Accord will have

beneficial effects on banks’ capital requirements (lower) linked to the SME segment,

either if the Standardized or one of the IRB (Advanced or Foundation) approaches is

used. The various studies find that it is easier to assess the amount of the savings in

banks’ capital requirements with the Standardized approach than with the IRB approach.

Using the Standardized approach to calculate capital requirements results in no savings if

the SMEs are considered as corporate (same 8% as before); but, if they are considered as

retail, the risk weight goes from 100% to a new 75% (Basel Committee on Banking

Supervision, June 2004, par. 69), so the capital requirement effectively becomes 6%. The

IRB approaches, however, allow banks to personalize the capital requirement

calculation,6 building their own models in order to estimate PDs (with the Foundation)

or even LGDs (with the Advanced) for each client. This means that it is not possible to

assess the exact savings in capital requirements for banks that will implement the IRB

approach, since these will be linked to the specific models utilized and to the

characteristics of each portfolio of credits. In our study, we actually do provide models

and realistic assumptions to perform this analysis.

Most studies have tried to understand the most probable risk weights and their possible

changes under Basel II using a generic sample of small and medium sized enterprises and

assumptions in order to estimate PD and LGD. Since we do not have the necessary data

to estimate LGDs, we are forced to use some realistic assumptions. We focus our

analysis on the probability of default, developing three specific credit risk models with

6 See Heitfield (2004) which explains how banks should choose their own Brating philosophy.^

EFFECTS OF THE NEW BASEL CAPITAL ACCORD 17

the purpose of estimating PDs in the most Basel II compliant way. Then, to test the

soundness of the assumptions that we made in order to derive capital requirements, we

provide a sensitivity analysis with respect to LGD and the asset’s maturity.

2.1. European literature

In a study of the Austrian economy, Schwaiger (2002) uses a sample of 11,610 enterprises,

with revenues between "1 and "50 million. A scoring system7 is applied in order to

assign a score to each enterprise and then these scores are grouped into twelve rating

classes. PDs for each rating class are calculated by dividing the number of defaults by the

number of enterprises in every class. A fixed LGD of 45% is used, based on the

assumption suggested in the Foundation IRB approach (F-IRB) for senior, unsecured

loan exposures.8 Loan exposures are approximated by enterprises’ revenues9 or by the

percentage of firms in each rating class. The formulas used to calculate the bank capital

requirements are the ones contained in the October 2002 version of the Accord (QIS 3).

The main difference with the final Basel II formulas (June 2004) is that expected losses

are not subtracted from the capital requirements. SMEs are considered only as corporate,

applying the size Bdiscount,^ and not as retail. Final results show a cumulative capital

requirement of 5.40%, if calculated by using the revenue assumption, or 6.36%, if the

weights for the different rating classes are approximated by the percentage of firms in

each class. Both results demonstrate that Basel II (QIS3) reduces quite substantially the

banks capital requirements for Austrian SMEs, compared to the current 8%. It is

important to point out that expected losses (PDxLGD) are not subtracted from the capital

requirements, as is required in the last version of the Accord.

For the Spanish economy, Saurina and Trucharte (2004) apply their analysis to a more

complex environment. Results are derived by using the aggregate average PDs for

various size categories for eight years (1994Y2001) of data, and not through a scoring

system, as the Basel Accord requires. The same formulas, as in the Austrian work, are

utilized and the expected losses (PDxLGD) are again considered as part of capital

requirements. Their first empirical result shows that the smaller an obligor (measured by

its volume of sales), the greater its probability of default.10 Indeed, almost all extant

literature agrees that there is no evidence that SME loans are less risky, in terms of lower

expected losses, than large corporate firms. The Basel Committee and others (e.g.( Lopez,

2004) state only that the asset correlation amongst firms is positively related to the size of

a firm and this explains the Bcorrelation discount^ for firms with less than "50 million of

sales. Moreover, this assumption is probably related to the possibility for banks to

7 The BBonitatsindex^ from Creditreform (Austria) is transformed into a rating system.

8 Basel Committee on Banking Supervision, QIS3, October 2002, par. 246.

9 Higher amounts are supposed to be granted to better quality enterprises and a similar distribution is found

for revenues.

10 There are many works on this topic that show the same results, especially for the US market, e.g.( Berger and

Udell (1990, 1995), Booth (1992). More recently a study by Jacobson et al. (2004) demonstrates, for two

Swedish commercial banks portfolios of loans, that the new Basel formulas are successful in covering the

actual loss rates for the SME portfolio, considered both as corporate or as retail.

18 ALTMAN AND SABATO

classify SMEs loans as retail. We will present our views about this subject at a later

point.

The other main result from the Spanish study concerns the amount of bank capital

requirements that are calculated using the IRB approach for large firms (6.50%), SMEs

corporate (8.94%) and SMEs retail (6.26%). The authors demonstrate that requirements

for SMEs considered as retail are considerably less than those for SMEs considered as

corporate. However, both are higher than those that would be obtained using the

Standardised approach, 8% (for SMEs classified as corporate) and 6% (for SMEs

classified as retail). We have to underline again that in the Spanish work no rating classes

or exposure weights are used when capital requirements are calculated with the IRB

approach and this could significantly affect results since the main goal of the Basel II

A-IRB method is the rating of assets in a portfolio based on their different risks and to

assign them diverse risk weights.

The European Council of Ministers meeting in Barcelona in March 2002, and sub-

sequently the European Parliament, requested the Commission Bto present a report on the

consequences of the Basel deliberations for all sectors of the European economy with

particular attention to SMEs.^ The report (PricewaterhouseCoopers and NIESR, 2004)

extensively analyzes European SMEs and their financial structure. Many academic

studies for different countries are compared in order to understand the most probable

effects of Basel II on credit risk capital requirements for SMEs. The main conclusions are

that the new Accord should not have any negative impact on the availability and cost of

finance for SMEs in most European countries. The study points out that worries about an

increase in the cost of finance due to an increased use of internal ratings in lending activity

are not justified. Indeed, they conclude that credit risk capital requirements relating to

SMEs are expected to decrease, especially if an IRB approach is used.

2.2. US literature

The United States position with respect to Basel II implementation is slightly more

complex than the European one and has been the source of many discussions. In August

2003, US banking regulators11 proposed a three-tier system in which only large,12

internationally active banking organizations (core banks) will be required to adopt the

A-IRB approach on a mandatory basis, while the other organizations, if subject to the

Accord,13 could choose to adopt that approach (opt-in banks) or not (general banks).

Most banking organizations will more than likely remain Basel I banks.

11 Advanced Notice of Proposed Rulemaking (ANPR) for the application of Basel II to US banking

organizations, August 2003.

12 Banking organizations with total banking (and thrift) assets of at least $250 billion or at least $10 billion in

on-balance-sheet foreign exposures. These kinds of banks are expected to number between 10 and 20 in

US market.

13 The 1988 Accord was applied only to internationally active banks. Under the new Accord, the scope of

application has been broadened also to encompass bank holding companies that are parents of

internationally active Bbanking groups.^

EFFECTS OF THE NEW BASEL CAPITAL ACCORD 19

Many aspects of this proposal have been analyzed in the academic literature (Allen

and Saunders, 2004; Hannan and Pilloff, 2004), but the work of Berger (2004)

specifically focuses on the effects of Basel II on banks in the US SME credit market.

He uses the results of the Quantitative Impact Study 3 (QIS3), carried out by the Basel

Committee in 2002 through a survey of large US banking organizations. Seventeen of the

22 responding institutions provided estimates of the A-IRB adoption effects on their

SME portfolios showing an average reduction of 33% in required credit risk capital. We

must underline again that the A-IRB formulas used at that time to calculate bank capital

requirements still considered expected loss as part of these requirements. This means, as

we show later, that a higher reduction could be expected applying the new formulas.

Based on these results, even if considered Brough estimates,^ and on other consid-

erations, the study concludes that the benefits of a decline in marginal costs of SME

lending for banking organizations that adopt the A-IRB approach are likely to only

slightly affect small banks which will not adopt that method, since they have a greater

competitive advantage given by their Brelationship lending.^ Otherwise, Basel II

implementation may significantly adversely influence the competitive positions in the

SME market of large banking organizations in the US that will not adopt the A-IRB,

since they share the same kind of Bhard^ information.14

3. Deriving bank capital requirements

In this section we explain how we derive bank capital requirements relating to SMEs for

the three countries that we consider. Methodologies, formulas and assumptions are

clearly defined in order to allow better understanding of the final results. We first

examine the definition of a SME comparing the position in different countries and the

one assumed by the Basel Committee.

3.1. SME definition

We find that there is no common definition of the segment of small and medium sized

enterprises across different countries. The definition varies from country to country,

taking into account different quantitative15 and qualitative16 variables. We restrict our

focus to three important economic zones (US, E.U. and Australia) that will be impacted

by the Basel Accord revisions.

14 Berger (2004) distinguishes between the kind of information available for large banking organizations

(Bhard^) and the one (Bsoft^) available for smaller banks (community banks).

15 The most commonly used are: annual turnover, total assets, number of employees, average annual receipts

or capital.

16 Usually no attention is given to the legal form, but independence from big enterprises, work organization

and industry type are often considered.

20 ALTMAN AND SABATO

The European Union has had a common definition since 1996 that was updated in

2003,17 probably to take into account also the new Basel rules. The number of employees

and the annual turnover of a firm are the criteria considered (less than "50 million in

sales or less than 250 employees).

In the United States there is a special organization (Small Business Administration, or

SBA) that deals with the politics relating to SMEs and also with their definition based on

the North American Industry Classification System (NAICS). Four criteria are used to

identify small business firms:18 three generic qualitative rules and one quantitative

requirement linked to the industry type. In general, the maximum number of employees

is 500 and the average annual receipts should be less than $28.5 million, but these limits

are different for each industry.

Australia has many definitions of SMEs linked to the context in which they are used.19

The most common is that the annual turnover should be less than $10 million, but

sometimes the number of employees (less than 50) is used.

Facing all these different criteria, the Basel Committee has mainly chosen to follow

the annual turnover definition, setting the same general rules to calculate the capital

requirements for all the firms (large, medium and small), but ensuring a lighter treatment

for small and medium ones (with annual turnover less than "50 million).

We believe that this decision, as already expressed, is based on the assumption that

these firms have a lower default correlation with each other and not because they are

considered less risky, in terms of lower expected losses, than the larger firms. Moreover, a

part of SMEs can be classified as retail, but in this case the SME definition does not play

any role. The only criterion considered is the bank’s exposure (less than "1 million). We

conclude that, with this rule, banks are motivated to utilize automatic decision systems to

manage clients with lower exposures, regardless of whether they are firms or private

individuals, in order to improve their internal efficiency. Scoring or rating systems, in the

application and the behavioral process, are the best way to treat Bhard information^ for

the retail segment where low profit margins and high volumes of customers do not allow an

efficient Brelationship lending^ (See Udell 2004, for further discussions).

3.2. The capital requirements

In Pillar 1 of the new Basel Capital Accord, the rules to calculate bank capital

requirements for each of the different segments, or Bbuckets,^ are clearly explained. We

focus on the formulas that can be used for SMEs when the A-IRB approach is applied.

As already stated, banks can classify SMEs as retail, utilizing the formula for Bother

17 Commission Recommendation 96/280/EC of April 3, 1996, updated in 2003/361/EC of May 6, 2003.

18 A small business is one that: 1) is organized for profit; 2) has a place of business in the US; 3) makes a

significant contribution to the US economy by paying taxes or using products, materials or labor; and

4) does not exceed the numerical size standard for its industry. For the specific table, see http://www.

sba.gov/size/summary-whatis.html.

19 See Gibson (2001).

EFFECTS OF THE NEW BASEL CAPITAL ACCORD 21

retail exposures,^ or as corporate, applying the corporate formula with the firm-size

adjustment.20 All formulas follow the same calculation steps involving inputs for

correlation (R), capital requirement (K) and risk-weighted assets (RWA).21 The most

important input variables, to be provided by the banks, are three (PDs, LGDs and

exposures at default (EADs)22), while the asset correlation (R) is implicitly given by the

Basel formulas. The exact formulas for SMEs from the final version of the Basel Capital

Accord are shown in table 1. Since, in our data samples, we do not have the loan exposures

or the possibility to estimate the LGD for each counterparty, we have to make some

assumptions, basically following the literature discussed in section 2.

A fixed LGD of 45% is assumed, using the one suggested in the Foundation IRB approach

(F-IRB) for senior, unsecured loan exposures,23 and the percentage of firms in each rating

class is used as weight for the capital requirement, instead of the loan exposure. We believe

that both assumptions are reasonable and do not have material effects on the final results.24

To be sure of this, we provide, in Appendix D, a sensitivity analysis to show the effects of

using different values of LGD and maturity. We focus on one year PDs estimation,

developing three statistical models in order to discriminate SME risk in each country sample.

Then, we point out that the lower correlation value for the retail assets, implicitly assumed by

the Basel formulas, is the main driver of the lower capital requirement for that segment.

Table 1. New SME capital requirement formulas

This table shows the new formulas, the ones contained in the last version (June 2004) of the Basel Accord. For

the SMEs classified as retail, the formulas are the ones for the Bother retail exposures.^ When SMEs are

classified as corporate, the formulas to be used are the one for the corporate, considering the size discount.

Source: Basel Committee on Banking Supervision 2004. BInternational Convergence of Capital Measurement

and Capital Standards.^

SME as retail SME as corporate

Correlation = R

= 0.03*(1 j EXP(j35*PD))/(1 j EXP(35))

+ 0.16*[1 j (1 j EXP(j35*PD))

/(1 j EXP(j35))]

Correlation = R. = 0.12*(1 j EXP(j50*PD))

/(1 j EXP(j50)) + 0.24*(1 j (1 j EXP

(j50*PD))/(1 j EXP(j50)))

j 0.04*(1 j (S j 5)/45)

Capital requirement = K

= LGD*N((1 j R)^ j 0.5)*G(PD)

+ (R/(1 j R)^0.5)*G(0.999)) j PD*LGD

Capital requirement = K

= (LGD*N((1 j R)^ j 0.5)*G(PD)

+ (R/(1 j R)^0.5)*G(0.999)) j PD*LGD)

*(1 j 1.5*b)^(j1*(1 + (M j 2.5)*b))

Maturity adjustment = (b)

= (0.11852 j 0.05478*LN(PD)^2)

20 Basel Committee on Banking Supervision, June 2004, par. 330 and par. 272Y273.

21 RWA = K � 12.5 � EAD. Since the final capital requirement is the 8% of this amount, we do not multiply

the RWA by 12.5.

22 As expressed in par. 334 of the Accord, we assume EAD for on-balance sheet items equal to the loan

exposure amount.

23 Basel Committee on Banking Supervision, June 2004, par. 287.

24 Perhaps we will observe a slight overestimate since many empirical studies show LGD distributions as

highly abnormal with a pronounced positive skew and also we should expect higher exposures for better

quality borrowers.

22 ALTMAN AND SABATO

Before explaining the models, it is important to discuss the variables used since,

amongst SMEs, there are different kinds of customers according to their legal form.

Basically, they could be sole traders, partnerships or limited companies, but the main

problem is that for each one of these possibilities a different set of information is

available. Specifically, for partnerships and limited companies, usually a great deal of

balance sheet variables is generally used to develop scoring models, as in our study, but

these are not available for sole traders. We believe that in this latter case, the only

possibility is to utilize personal information about the owners, considering them as

private individuals. For example, internal sources, such as the client’s behavior with

other bank products, or external sources, such as public or credit bureau data, could be

used. Moreover, in order to improve the performance of the models for all the possible

kinds of firms, generic firm quantitative and qualitative variables25 could be very useful.

In our data samples, we exclude sole traders, since we do not have enough variables to

develop a specific model for them, while, for other firms, we consider only financial

statement variables (financial ratios).

3.2.1. The Italian case. In the Italian sample, there are 20,193 SME firms derived

from a portfolio of a large Italian bank. We know that none of them have sales over "50

million and an exposure over "1 million, but we do not have the exact amounts for each

firm. About 40 financial ratios have been analyzed and 20 of them have been chosen to

run in a logistic regression to develop our model (See Appendix A for a detailed

discussion about how the model has been developed). The final model contains eight

variables: three describing the leverage structure, one the profitability and four the

liquidity situation of the firm. Our model creates nine rating classes that allow

discriminating PDs ranging from 0.03% to 15%. For each rating class, PDs are

calculated by dividing the number of defaults by the number of enterprises in each class.

Rating classes have been created in order to obtain the value of PD closest to the one

showed by bond equivalent PD distributions (Column 2 in table 2). Classifying all SMEs

as retail, we obtain a capital requirement of 4.88% for Italy (table 2).

To consider SMEs as corporate (table 3), we have to make two additional

assumptions. The first is the effective maturity (Meff.), three years for smaller firms and

five years for medium sized firms. The maturity adjustment [(b)corp.] is a function only of

PD. The second assumption is about the amount of sales to use for the size adjustment.

We split the SME population into two groups: one with sales between "5 and "25 million

(small) and the other with sales between "26 and "50 million (medium). In this way we

use an average amount of sales of "10 and "30 million in each group, respectively.26 The

two groups’ percentages of capital requirements are aggregated, considering the most

likely distribution of these size firms in banks’ portfolios. We assume that the SME

25 Such as the number of employees, the legal form of the business, the region where the main business is

carried out, the industry type, etc. See Grunert et al. (2004), for further discussions about the role of non-

financial factors in internal credit ratings.

26 These average amounts are based on the realistic assumption that each sales distribution is skewed with

more relatively small borrowers than relatively large borrowers.

EFFECTS OF THE NEW BASEL CAPITAL ACCORD 23

portfolios are typically comprised of 85% of small and 15% of medium sized firms.27

The resulting weighted capital requirement of the two size components of SMEs,

calculated in table 3, is 8.45% (0.85*8.25 + 0.15*9.66) for the Italian SME population.

3.2.2. The Australian case. For the Australian case, we apply the Corporate Score-

card Credit Rating (CSCR), derived by the Altman et al. (1977) ZETA score model,

adopted for Australia by the Corporate Scorecard Group. We obtain 15 rating classes,

with expected PDs from 0.03% to 20%. The assumptions and considerations are the same

made for Italy. The PDs are derived from the Basel Accord’s directives for each bond

equivalent rating class.

The Australian sample contains 10,000 firms, 5,749 of which are SMEs, with sales less

than "50 million, that can be classified as retail or corporate. When all SMEs are considered

as retail, results show a capital requirement of 4.62% (table 4). For SMEs classified as

corporate (table 5), we split the population into two groups28 (small and medium) using the

same average amount of sales ("10 and "30 million). The resulting weighted capital

requirement for Australian corporate SMEs is 8.81% (0.64*8.33 + 0.36*9.71).

We also calculate the capital requirement for 4,251 large companies, with sales over

"50 million. The result is a cumulated capital requirement of 7.83% (table 6), just

slightly less than the current 8% requirement.

Table 2. All SMEs as retail (Italy)

This table shows the capital requirement when all SMEs are classified as retail. In the first column, rating

classes have been assigned on a bond equivalent basis. In the second and third columns, probability of default

(PD) and loss given default (LGD) are shown. In the fourth column, the correlation for each rating class is

calculated using the formula: Rsme = Correlation = 0.03*(1 j EXP(j35*PD))/(1 j EXP(j35)) + 0.16*[1 j

(1 j EXP(j35*PD))/(1 j EXP(j35))]. In the fifth column, the capital requirement linked to each rating class

is calculated with the following formula: Ksme = Capital requirement = LGD*N((1 j R)^ j 0.5)*G(PD) +

(R/(1 j R)^0.5)*G(0.999)) j PD*LGD. In the sixth column, the weights are assigned utilizing the percent of

firms in each rating class. In the last column, the product of the capital requirement (Ksme) and the weight is

cumulated to obtain the total capital requirement.

Rating PD LGD Rsme Ksme Weight Cum. weighted Ksme

AAA 0.03% 45% 0.15864 0.0035609 0.0000 0.000%

AA 0.05% 45% 0.15774 0.0053033 0.0045 0.002%

BBB 0.44% 45% 0.14145 0.0240566 0.0226 0.057%

BB 1.06% 45% 0.11971 0.0375285 0.0746 0.337%

BBj 1.91% 45% 0.09662 0.0458465 0.2622 1.539%

B+ 3.31% 45% 0.07081 0.0509144 0.3578 3.361%

B 5.61% 45% 0.04825 0.0537624 0.2551 4.732%

Bj 9.34% 45% 0.03495 0.0591754 0.0143 4.817%

CCC 15.00% 45% 0.03068 0.0708806 0.0089 4.880%

27 This breakdown was suggested to us by Centrale dei Bilanci, an Italian organization based in Turin, owned

by over fifty banks and the Bank of Italy, which has access to the financial statements of more than 40,000

Italian firms.

28 This breakdown was suggested to us by the Corporate Scorecard Group, the company that provided us the

data. In the sample, out of 5,749 SMEs, 3,650 (64%) were small and 2,099 (36%) were medium sized.

24 ALTMAN AND SABATO

3.2.3. The United States case. For the United States, we use data from WRDS

Compustat29 for US firms during the period 2000Y2003. Considering the variables

contained in the original Z-score model for manufacturing and non-manufacturing firms

(Altman, 1993),30 a new model has been developed specifically for SMEs. The actual

model’s construction is given in Appendix B. Using this model, we create eleven rating

classes that allow discriminating PDs from 0.03% to almost 20%. In order to calculate

the capital requirements, the same assumptions as in the previous cases are made.

The SME sample contains 3,552 firms, 2,730 (77%) small (with sales less than "25

million) and 822 (23%) medium sized (with sales between "26 and "50 million). The

Table 3. All SMEs as corporate (Italy)

This table shows the capital requirement when all SMEs are classified as corporate. In the first column, rating classes

have been assigned on a bond equivalent basis. In the second and third column, probability of default (PD) and loss

given default (LGD) are shown. In the fourth column, the correlation for each rating class is calculated using the

formula: Rcorp. = Correlation = 0.12*(1 j EXP(j50*PD))/(1 j EXP(j50)) + 0.24*(1 j (1 j EXP(j50*PD))/

(1 j EXP(j50))) j 0.04*(1 j (S j 5)/45), where S is the amount of sales for each firm. In the fifth column, the

maturity adjustment is calculated as: (b)corp. = Maturity adjustment = (0.11852 j 0.05478*LN(PD)^2). In the sixth

column there is the effective maturity (Meff). In the seventh column, the capital requirement linked to each rating

class is calculated with the following formula: Kcorp. = Capital requirement = (LGD*N((1 j R)^ j 0.5)*G(PD) +

(R/(1 j R)^0.5)*G(0.999)) j PD*LGD)*(1 j 1.5*b)^(j1*(1 + (M j 2.5)*b)). In the eighth column the weights

are assigned utilizing the percent of firms in each rating class. In the last column the product of the capital

requirement (Ksme) and the weight is cumulated to obtain the total capital requirement.

Rating PD LGD Rcorp. ( b)corp. Meff. Kcorp. Weight Cum. weighted Kcorp.

Sales "5Y25 mil

AAA 0.03% 45% 0.20266 0.31683 3.0 0.0103082 0.0000 0.000%

AA 0.05% 45% 0.20148 0.28612 3.0 0.0137376 0.0045 0.006%

BBB 0.44% 45% 0.18075 0.17286 3.0 0.0437287 0.0226 0.105%

BB 1.06% 45% 0.15508 0.13513 3.0 0.0617752 0.0746 0.566%

BBj 1.91% 45% 0.13062 0.11245 3.0 0.0730465 0.2622 2.481%

B+ 3.31% 45% 0.10738 0.09316 3.0 0.0834057 0.3578 5.465%

B 5.61% 45% 0.09171 0.07635 3.0 0.0973451 0.2551 7.949%

Bj 9.34% 45% 0.08557 0.06170 3.0 0.1185345 0.0143 8.118%

CCC 15.00% 45% 0.08451 0.04948 3.0 0.1425111 0.0089 8.245%

Sales "26Y50 mil

AAA 0.03% 45% 0.22044 0.31683 5.0 0.0173454 0.0000 0.000%

AA 0.05% 45% 0.21926 0.28612 5.0 0.0211578 0.0045 0.010%

BBB 0.44% 45% 0.19852 0.17286 5.0 0.0540392 0.0226 0.132%

BB 1.06% 45% 0.17285 0.13513 5.0 0.0734324 0.0746 0.679%

BBj 1.91% 45% 0.14840 0.11245 5.0 0.0857668 0.2622 2.928%

B+ 3.31% 45% 0.12515 0.09316 5.0 0.0975877 0.3578 6.420%

B 5.61% 45% 0.10948 0.07635 5.0 0.1135301 0.2551 9.316%

Bj 9.34% 45% 0.10335 0.06170 5.0 0.1367453 0.0143 9.512%

CCC 15.00% 45% 0.10229 0.04948 5.0 0.1618240 0.0089 9.656%

29 Compustat North America (Standard & Poor’s Corp., a division of Mc Graw-Hill Corp.) is a database of

US and Canadian fundamental and market information on more than 24,000 active and inactive publicly

held companies from 1994 to 2003.

30 See also Altman et al. (1995).

EFFECTS OF THE NEW BASEL CAPITAL ACCORD 25

capital requirement is calculated based on whether the SMEs are considered all as retail

or all as corporate. For the retail case (table 7), the capital requirement is 4.65%. When

all SMEs are classified as corporate (table 8), the resulting weighted capital requirement

for the US SME population is 8.61% (0.77*8.29 + 0.23*9.67).

For large corporate firms, where sales are over "50 million, the original Z-Score model

for manufacturing and non-manufacturing firms (Altman, 1993) has been applied to over

3,700 US companies extracted from the WRDS Compustat database for the years

2000Y2003. Seven rating classes have been created, discriminating PDs from 0.03% to

20%. Results show a capital requirement of 6.64%, consistently lower than the current

8% requirement (table 9).

3.3. Summary of results

In table 10, we summarize the bank capital requirements that we believe are likely under

Basel II for all three of our countries. The results are consistent with our expectations.31

Table 4. All SMEs as retail (Australia)

This table shows the capital requirement when all SMEs are classified as retail. In the first column, rating classes are

listed. In the second and third columns, probability of default (PD), assigned on a bond equivalent basis, and loss given

default (LGD) are shown. In the fourth column, the correlation for each rating class is calculated using the formula:

Rsme = Correlation = 0.03*(1 j EXP(j35*PD))/(1 j EXP(j35)) + 0.16*[1 j (1 j EXP(j35*PD))/(1 j

EXP(j35))]. In the fifth column, the capital requirement linked to each rating class is calculated with the following

formula: Ksme = Capital requirement = LGD*N((1 j R)^ j 0.5)*G(PD) + (R/(1 j R)^0.5)*G(0.999)) j PD*LGD.

In the sixth column, the weights are assigned utilizing the percent of firms in each rating class. In the last column, the

product of the capital requirement (Ksme) and the weight is cumulated to obtain the total capital requirement.

Rating PD LGD Rsme Ksme Weight Cum. weighted Ksme

AAA 0.03% 45% 0.15864 0.0035609 0.0009 0.000%

AA 0.07% 45% 0.15685 0.0068515 0.0236 0.016%

A+ 0.09% 45% 0.15597 0.0082644 0.0091 0.024%

A 0.10% 45% 0.15553 0.0089303 0.0182 0.040%

Aj 0.11% 45% 0.15509 0.0095731 0.0318 0.071%

BBB+ 0.20% 45% 0.15121 0.0145806 0.0300 0.114%

BBB 0.30% 45% 0.14704 0.0190389 0.0381 0.187%

BBBj 0.50% 45% 0.13913 0.0258890 0.0390 0.288%

BB+ 0.75% 45% 0.12999 0.0320784 0.0572 0.471%

BB 1.00% 45% 0.12161 0.0366182 0.0744 0.744%

BBj 1.50% 45% 0.10690 0.0426968 0.1361 1.325%

B+ 2.60% 45% 0.08233 0.0490728 0.1742 2.180%

B 6.00% 45% 0.04592 0.0541848 0.1416 2.947%

Bj 10.00% 45% 0.03393 0.0604342 0.0717 3.380%

CCC 20.00% 45% 0.03012 0.0802219 0.1543 4.618%

31 Results for Italy, especially for SMEs considered as retail, are very close to those reported in a research

made by Prometeia and the University of Bologna [see Bocchi and Lusignani (2004)]. They applied the

A-IRB on 75,000 Italian SMEs obtaining a capital requirement of 4.8% (SMEs retail) and 7.5% (SMEs

corporate).

26 ALTMAN AND SABATO

Table 5. All SMEs as corporate (Australia)

This table shows the capital requirement when all SMEs are classified as corporate. In the first column, rating

classes are listed. In the second and third columns, probability of default (PD), assigned on the bond equivalent

basis, and loss given default (LGD) are shown. In the fourth columns, the correlation for each rating class is

calculated using the formula: Rcorp. = Correlation = 0.12*(1 j EXP(j50*PD))/(1 j EXP(j50)) + 0.24*(1 j

(1 j EXP(j50*PD))/(1 j EXP(j50))) j 0.04*(1 j (S j 5)/45), where S is the amount of sales for each firm.

In the fifth column, the maturity adjustment is calculated as: ( b)corp. = Maturity adjustment = (0.11852 j

0.05478*LN(PD)^2). In the sixth column there is the effective maturity (Meff). In the seventh column, the

capital requirement linked to each rating class is calculated with the following formula: Kcorp. = Capital

requirement = (LGD*N((1 j R)^ j 0.5)*G(PD) + ( R/(1 j R)^0.5)*G(0.999)) j PD*LGD)*(1 j

1.5*b)^(j1*(1 + (M j 2.5)*b)). In the eighth column the weights are assigned utilizing the percent of firms

in each rating class. In the last column the product of the capital requirement (Ksme) and the weight is

cumulated to obtain the total capital requirement.

Rating PD LGD Rcorp. (b)corp. Meff. Kcorp. Weight Cum. weighted Kcorp.

Sales "5Y25 mil

AAA 0.03% 45% 0.20266 0.31683 3.0 0.0103082 0.0009 0.001%

AA 0.07% 45% 0.20032 0.26674 3.0 0.0166174 0.0236 0.040%

A+ 0.09% 45% 0.19916 0.25271 3.0 0.0191437 0.0091 0.058%

A 0.10% 45% 0.19859 0.24694 3.0 0.0203066 0.0182 0.095%

Aj 0.11% 45% 0.19802 0.24177 3.0 0.0214139 0.0318 0.163%

BBB+ 0.20% 45% 0.19302 0.21064 3.0 0.0296251 0.0300 0.251%

BBB 0.30% 45% 0.18773 0.19075 3.0 0.0364547 0.0381 0.390%

BBBj 0.50% 45% 0.17790 0.16709 3.0 0.0462932 0.0390 0.571%

BB+ 0.75% 45% 0.16692 0.14942 3.0 0.0546619 0.0572 0.884%

BB 1.00% 45% 0.15723 0.13749 3.0 0.0605946 0.0744 1.334%

BBj 1.50% 45% 0.14113 0.12151 3.0 0.0685692 0.1361 2.268%

B+ 2.60% 45% 0.11715 0.10141 3.0 0.0786754 0.1742 3.638%

B 6.00% 45% 0.09042 0.07433 3.0 0.0996762 0.1416 5.050%

Bj 10.00% 45% 0.08525 0.05986 3.0 0.1218895 0.0717 5.924%

CCC 20.00% 45% 0.08445 0.04272 3.0 0.1557644 0.1543 8.327%

Sales "26Y50 mil

AAA 0.03% 45% 0.22044 0.31683 5.0 0.0173454 0.0009 0.002%

AA 0.07% 45% 0.21809 0.26674 5.0 0.0243731 0.0236 0.059%

A+ 0.09% 45% 0.21694 0.25271 5.0 0.0271871 0.0091 0.084%

A 0.10% 45% 0.21637 0.24694 5.0 0.0284790 0.0182 0.136%

Aj 0.11% 45% 0.21580 0.24177 5.0 0.0297070 0.0318 0.230%

BBB+ 0.20% 45% 0.21080 0.21064 5.0 0.0387449 0.0300 0.346%

BBB 0.30% 45% 0.20551 0.19075 5.0 0.0461809 0.0381 0.522%

BBBj 0.50% 45% 0.19568 0.16709 5.0 0.0567986 0.0390 0.744%

BB+ 0.75% 45% 0.18470 0.14942 5.0 0.0657834 0.0572 1.120%

BB 1.00% 45% 0.17501 0.13749 5.0 0.0721596 0.0744 1.657%

BBj 1.50% 45% 0.15891 0.12151 5.0 0.0808140 0.1361 2.757%

B+ 2.60% 45% 0.13493 0.10141 5.0 0.0921371 0.1742 4.362%

B 6.00% 45% 0.10820 0.07433 5.0 0.1161421 0.1416 6.006%

Bj 10.00% 45% 0.10303 0.05986 5.0 0.1403267 0.0717 7.013%

CCC 20.00% 45% 0.10223 0.04272 5.0 0.1750002 0.1543 9.713%

EFFECTS OF THE NEW BASEL CAPITAL ACCORD 27

Capital requirements under the A-IRB approach, if all SMEs will be classified as retail,

should be between 1 and 2 percentage points less than the requirement (6%) if the

Standardized approach would be used. Many factors will play a role in assessing the

exact savings, but the primary determinants will be the quality of the SME portfolio and

of the models used to assign PDs and LGDs. Moreover, the difference compared to the

current requirement (8%) is likely to be lower by about 3 percentage points or more. On

the other hand, if all SMEs should be considered as corporate, the requirements will

likely be higher than the current 8%. We believe that many banks will be obliged to

choose a blended approach (considering some SMEs as retail and some as corporate).

Based only on credit risk considerations, our breakeven analysis shows that, for all of

our countries, banking organizations should classify as retail no less than the 20% of

their SME portfolio in order to not increase their current capital requirement (8%)

relating to SMEs. However, we believe that this percentage is likely to be higher if other

kinds of risk, for example operational risk, were considered in the capital requirement

calculation.

Looking at the entire set of results, we can conclude that concerns about an increase

in capital requirements are not justified for the SME portfolio, especially if banks will

have the chance to classify at least the 20% as retail. Moreover, our breakeven analysis

Table 6. Corporate—Large firms (Australia)

This table shows the capital requirement for large corporate firms. In the first column, rating classes are listed.

In the second and third columns, probability of default (PD), assigned on a bond equivalent basis, and loss

given default ( LGD) are shown. In the fourth column, the correlation for each rating class is calculated using

the formula: Rcorp. = Correlation = 0.12*(1 j EXP(j50*PD))/(1 j EXP(j50)) + 0.24*(1 j (1 j

EXP(j50*PD))/(1 j EXP(j50))). In the fifth column, the maturity adjustment is calculated as: ( b)corp. =

Maturity adjustment = (0.11852 j 0.05478*LN(PD)^2). In the sixth column there is the effective maturity

(Meff). In the seventh column, the capital requirement linked to each rating class is calculated with the

following formula: Kcorp. = Capital requirement = ( LGD*N((1 j R)^ j 0.5)*G(PD) + ( R/(1 j

R)^0.5)*G(0.999)) j PD*LGD)*(1 j 1.5*b)^(j1*(1 + (M j 2.5)*b)). In the eighth column the weights are

assigned utilizing the percent of firms in each rating class. In the last column the product of the capital

requirement (Ksme) and the weight is cumulated to obtain the total capital requirement.

Rating PD LGD Rcorp. (b)corp. Meff. Kcorp. Weight Cum. weighted Kcorp.

AAA 0.03% 45% 0.23821 0.31683 5.0 0.0192569 0.0000 0.000%

AA 0.07% 45% 0.23587 0.26674 5.0 0.0270318 0.0235 0.064%

A+ 0.09% 45% 0.23472 0.25271 5.0 0.0301399 0.0165 0.113%

A 0.10% 45% 0.23415 0.24694 5.0 0.0315661 0.0329 0.217%

Aj 0.11% 45% 0.23358 0.24177 5.0 0.0329213 0.0376 0.341%

BBB+ 0.20% 45% 0.22858 0.21064 5.0 0.0428891 0.0494 0.553%

BBB 0.30% 45% 0.22328 0.19075 5.0 0.0510895 0.0659 0.890%

BBBj 0.50% 45% 0.21346 0.16709 5.0 0.0628223 0.1341 1.732%

BB+ 0.75% 45% 0.20247 0.14942 5.0 0.0728073 0.1059 2.503%

BB 1.00% 45% 0.19278 0.13749 5.0 0.0799553 0.1341 3.575%

BBj 1.50% 45% 0.17668 0.12151 5.0 0.0897988 0.1765 5.160%

B+ 2.60% 45% 0.15270 0.10141 5.0 0.1029990 0.1412 6.614%

B 6.00% 45% 0.12597 0.07433 5.0 0.1305268 0.0494 7.259%

Bj 10.00% 45% 0.12081 0.05986 5.0 0.1566594 0.0188 7.554%

CCC 20.00% 45% 0.12001 0.04272 5.0 0.1919380 0.0141 7.825%

28 ALTMAN AND SABATO

(table 11) shows that by implementing the A-IRB instead of the Standardized approach,

at least 40% of SME portfolio should be considered as retail. For this to occur, banks

must consider not only the exposure requirement (less than "1 million) to classify SMEs

as retail, but they should also treat SMEs as Bretail^ in a pooling of assets.

This means that banks will be forced to update their systems and procedures to manage

SMEs in a more efficient way. Automated decision systems, such as scoring or rating

models, commonly used only for private individuals, are needed to evaluate Bhard

information^ about SMEs in a more profitable and efficient way.32 Even if, in the

short term, A-IRB adopting banks will likely face higher implementation costs, we

believe that in the long run they will, in fact, maximize their returns by reducing trans-

action costs.

Many recent studies33 find that smaller banks specialized in small business lending

enjoy a competitive advantage, mainly due to their ability to reduce default risk through

their Brelationship lending.^ We believe that with the new Basel Capital Accord, large,

well-diversified banks, which will adopt the A-IRB approach, will have the opportunity

to reap benefits in terms of either lower capital requirements or internal systems and

procedures improvement. Following the right strategies in credit risk management and

capital adequacy, large banking organizations are likely to reduce, or possibly eliminate,

the current competitive gap with small, specialized banks in the so-called middle market,

since smaller banks are more likely to use the higher capital requirements’ Standardized

Table 7. All SMEs as retail (US)

This table shows the capital requirement when all SMEs are classified as retail. In the first column, rating

classes are listed. In the second and third columns, probability of default (PD), assigned on a bond equivalent

basis, and loss given default (LGD) are shown. In the fourth column, the correlation for each rating class is

calculated using the formula: Rsme = Correlation = 0.03*(1 j EXP(j35*PD))/(1 j EXP(j35)) + 0.16*[1 j

(1 j EXP(j35*PD))/(1 j EXP(j35))]. In the fifth column, the capital requirement linked to each rating class

is calculated with the following formula: Ksme = Capital requirement = LGD*N((1 j R)^ j 0.5)*G(PD) + ( R/

(1 j R)^0.5)*G(0.999)) j PD*LGD. In the sixth column, the weights are assigned utilizing the percent of

firms in each rating class. In the last column, the product of the capital requirement ( Ksme) and the weight is

cumulated to obtain the total capital requirement.

Rating PD LGD Rsme Ksme Weight Cum. weighted Ksme

AAA 0.03% 45% 0.15864 0.0035609 0.0080 0.003%

AA 0.07% 45% 0.15685 0.0068515 0.0206 0.017%

A 0.12% 45% 0.15465 0.0101949 0.0319 0.049%

BBB+ 0.28% 45% 0.14786 0.0182169 0.0408 0.124%

BBB 0.35% 45% 0.14501 0.0209678 0.0612 0.252%

BB 1.26% 45% 0.11364 0.0401706 0.1975 1.046%

BBj 1.61% 45% 0.10400 0.0436673 0.1641 1.762%

B+ 2.75% 45% 0.07965 0.0495529 0.1437 2.474%

B 6.32% 45% 0.04423 0.0545513 0.1229 3.145%

Bj 10.34% 45% 0.03349 0.0611066 0.0920 3.707%

CCC 19.87% 45% 0.03012 0.0800075 0.1173 4.645%

32 See Berger and Udell (2004) about the role of lending infrastructures for SMEs.

33 See Kolari and Shin (2004) and Berger (2004).

EFFECTS OF THE NEW BASEL CAPITAL ACCORD 29

approach or to remain Basel I banks (in the U.S.). Moreover, the expected excess of

regulatory capital that would be created at A-IRB organizations as a result of reduced

capital requirements can motivate an increase in the acquisition process toward non-

adopting banking organizations.34 We also expect that the acquisitions of small US banks

will accelerate since the assets of these institutions will be more valuable on the balance

sheets of Basel II banks versus on Basel I banks’ balance sheets.

Table 8. All SMEs as corporate (US)

This table shows the capital requirement when all SMEs are classified as corporate. In the first column,

rating classes are listed. In the second and third columns, probability of default (PD), assigned on a bond

equivalent basis, and loss given default (LGD) are shown. In the fourth column, the correlation for each rating class

is calculated using the formula: Rcorp. = Correlation = 0.12*(1 j EXP(j50*PD))/(1 j EXP(j50)) + 0.24*(1 j

(1 j EXP(j50*PD))/(1 j EXP(j50))) j 0.04*(1 j (S j 5)/45), where S is the amount of sales for each firm.

In the fifth column, the maturity adjustment is calculated as: (b)corp. = Maturity adjustment = (0.11852 j

0.05478*LN(PD)^2). In the sixth column there is the effective maturity (Meff). In the seventh column, the

capital requirement linked to each rating class is calculated with the following formula: Kcorp. = Capital

requirement = (LGD*N((1 j R)^ j 0.5)*G( PD) + ( R/(1 j R)^0.5)*G(0.999)) j PD*LGD)*(1 j

1.5*b)^(j1*(1 + (M j 2.5)*b)). In the eighth column the weights are assigned utilizing the percent of firms in

each rating class. In the last column the product of the capital requirement (Ksme) and the weight is cumulated to

obtain the total capital requirement.

Rating PD LGD Rcorp. (b)corp. Meff. Kcorp. Weight Cum. weighted Kcorp.

Sales "26Y50 mil

AAA 0.03% 45% 0.20266 0.31683 3.0 0.0103082 0.0080 0.008%

AA 0.07% 45% 0.20032 0.26674 3.0 0.0166174 0.0206 0.042%

A 0.12% 45% 0.19746 0.23711 3.0 0.0224718 0.0319 0.114%

BBB+ 0.28% 45% 0.18877 0.19406 3.0 0.0352240 0.0408 0.258%

BBB 0.35% 45% 0.18518 0.18344 3.0 0.0392977 0.0612 0.498%

BB 1.26% 45% 0.14836 0.12826 3.0 0.0652130 0.1975 1.786%

BBj 1.61% 45% 0.13809 0.11882 3.0 0.0698990 0.1641 2.933%

B+ 2.75% 45% 0.11479 0.09946 3.0 0.0797299 0.1437 4.079%

B 6.32% 45% 0.08954 0.07279 3.0 0.1015793 0.1229 5.328%

Bj 10.34% 45% 0.08513 0.05896 3.0 0.1235601 0.0920 6.464%

CCC 19.87% 45% 0.08445 0.04287 3.0 0.1554966 0.1173 8.288%

Sales "26Y50 mil

AAA 0.03% 45% 0.22044 0.31683 5.0 0.0173454 0.0080 0.014%

AA 0.07% 45% 0.21809 0.26674 5.0 0.0243731 0.0206 0.064%

A 0.12% 45% 0.21523 0.23711 5.0 0.0308780 0.0319 0.163%

BBB+ 0.28% 45% 0.20655 0.19406 5.0 0.0448456 0.0408 0.346%

BBB 0.35% 45% 0.20296 0.18344 5.0 0.0492587 0.0612 0.647%

BB 1.26% 45% 0.16613 0.12826 5.0 0.0771527 0.1975 2.171%

BBj 1.61% 45% 0.15587 0.11882 5.0 0.0822757 0.1641 3.521%

B+ 2.75% 45% 0.13256 0.09946 5.0 0.0933470 0.1437 4.862%

B 6.32% 45% 0.10731 0.07279 5.0 0.1182625 0.1229 6.316%

Bj 10.34% 45% 0.10290 0.05896 5.0 0.1421017 0.0920 7.623%

CCC 19.87% 45% 0.10223 0.04287 5.0 0.1747425 0.1173 9.673%

34 See Hannan and Pilloff (2004).

30 ALTMAN AND SABATO

For SMEs, we expect an increase in lending activity, especially by large banking

organizations that will find this business more profitable. Concerns have been raised that

costs of the A-IRB implementation will likely increase credit prices for small and

medium sized enterprises. We suppose that this could be possible, at least at the

beginning, but, thanks to the new techniques for credit risk discrimination and to the

most advanced risk-based pricing methods, good quality firms will likely enjoy benefits

in terms of lower credit prices. The costs of transactions lending technologies will be

recovered in a few years considering that the possible expansion of SME business will

lead to better capital allocations.

Lastly, although we should be cautious with regard to the conclusions of this work,

since this paper has not addressed some other elements of the new Basel Accord (e.g., the

costs of operational risk), we conclude that the relationship between banks and SMEs

will more than likely change in a positive way for both.

Table 9. Corporate—Large firms (US)

This table shows the capital requirement for large corporate firms. In the first column, rating classes are listed.

In the second and third columns, probability of default (PD), assigned on a bond equivalent basis, and loss given

default (LGD) are shown. In the fourth column, the correlation for each rating class is calculated using the formula:

Rcorp. = Correlation = 0.12*(1 j EXP(j50*PD))/(1 j EXP(j50)) + 0.24*(1 j (1 j EXP(j50*PD))/(1 j

EXP(j50))). In the fifth column, the maturity adjustment is calculated as: (b)corp. = Maturity adjustment =

(0.11852 j 0.05478*LN(PD)^2). In the sixth column there is the effective maturity (Meff). In the seventh column,

the capital requirement linked to each rating class is calculated with the following formula: Kcorp. = Capital

requirement = (LGD*N((1 j R)^ j 0.5)*G(PD) + (R/(1 j R)^0.5)*G(0.999)) j PD*LGD)*(1 j 1.5*b)^(j1*

(1 + (M j 2.5)*b)). In the eighth column the weights are assigned utilizing the percent of firms in each rating

class. In the last column the product of the capital requirement (Ksme) and the weight is cumulated to obtain the

total capital requirement.

Rating PD LGD Rcorp. ( b)corp. Meff. Kcorp. Weight Cum. weighted Kcorp.

AAA 0.03% 45% 0.23821 0.31683 5.0 0.0192569 0.0381 0.073%

AA 0.07% 45% 0.23587 0.26674 5.0 0.0270318 0.1356 0.440%

A 0.10% 45% 0.23415 0.24694 5.0 0.0315661 0.2410 1.201%

BBB 0.30% 45% 0.22328 0.19075 5.0 0.0510895 0.2644 2.552%

BB 1.00% 45% 0.19278 0.13749 5.0 0.0799553 0.1288 3.582%

B 6.00% 45% 0.12597 0.07433 5.0 0.1305268 0.1016 4.908%

CCC 20.00% 45% 0.12001 0.04272 5.0 0.1919380 0.0904 6.643%

Table 10. Impact of Basel II on SMEs (Summary Results)

This table shows the capital requirements for all of our countries when the A-IRB approach is applied. In each

column, all of the countries are listed. For each row, the different possibilities in considering all SMEs as retail

or as corporate are shown. In the last row, the requirements for the large corporate segment are presented.

Italy US Australia

SMEs as retail 4.88% 4.65% 4.62%

SMEs as corporate 8.45% 8.61% 8.81%

CorporateVLarge firms n.a. 6.64% 7.83%

EFFECTS OF THE NEW BASEL CAPITAL ACCORD 31

Ta

ble

11

.B

rea

kev

ena

na

lysi

sfo

rca

pit

al

req

uir

emen

tso

fp

oss

ible

new

ap

pro

ach

esv

ersu

sth

ecu

rren

tsy

stem

Th

ista

ble

com

par

esth

ed

iffe

ren

tle

vel

so

fca

pit

alre

qu

irem

ents

for

each

app

roac

h,

con

sid

erin

gth

ep

erce

nta

ge

of

SM

Es

that

the

ban

kw

ill

wan

to

rw

ill

be

able

to

clas

sify

asre

tail

and

asco

rpo

rate

.F

or

all

of

ou

rco

un

trie

s,in

the

firs

tco

lum

n,

the

A-I

RB

req

uir

emen

tsar

ed

eriv

edb

yu

sin

gth

ere

sult

ssh

ow

nin

tab

le1

0m

ult

ipli

ed

by

the

per

cen

tag

eso

ffi

rms

con

sid

ered

asre

tail

and

asco

rpo

rate

.In

the

seco

nd

colu

mn

,th

eca

pit

alre

qu

irem

ent

isca

lcu

late

dm

ult

iply

ing

the

8%

req

uir

emen

tb

yth

e

per

cen

tag

eo

fS

ME

sco

nsi

der

edas

corp

ora

tep

lus

the

6%

req

uir

emen

tb

yth

ep

erce

nta

ge

of

SM

Es

con

sid

ered

asre

tail

.T

he

curr

ent

req

uir

emen

tis

alw

ays

8%

.

Per

cen

tag

eo

fS

ME

scl

assi

fied

asre

tail

and

asco

rpo

rate

Cap

ital

req

uir

emen

ts

Ital

yU

nit

edS

tate

sA

ust

rali

a

A-I

RB

Sta

nd

ard

ized

Cu

rren

tA

-IR

BS

tan

dar

diz

edC

urr

ent

A-I

RB

Sta

nd

ard

ized

Cu

rren

t

0%

SM

Es

asre

tail

10

0%

SM

Es

asco

rpo

rate

8.4

5%

8.0

0%

8.0

0%

8.6

1%

8.0

0%

8.0

0%

8.8

1%

8.0

0%

8.0

0%

10

%S

ME

sas

reta

il9

0%

SM

Es

asco

rpo

rate

8.0

9%

7.8

0%

8.0

0%

8.2

1%

7.8

0%

8.0

0%

8.3

9%

7.8

0%

8.0

0%

20

%S

ME

sas

reta

il8

0%

SM

Es

asco

rpo

rate

7.7

4%

7.6

0%

8.0

0%

7.8

2%

7.6

0%

8.0

0%

7.9

7%

7.6

0%

8.0

0%

30

%S

ME

sas

reta

il7

0%

SM

Es

asco

rpo

rate

7.3

8%

7.4

0%

8.0

0%

7.4

2%

7.4

0%

8.0

0%

7.5

5%

7.4

0%

8.0

0%

40

%S

ME

sas

reta

il6

0%

SM

Es

asco

rpo

rate

7.0

2%

7.2

0%

8.0

0%

7.0

3%

7.2

0%

8.0

0%

7.1

3%

7.2

0%

8.0

0%

50

%S

ME

sas

reta

il5

0%

SM

Es

asco

rpo

rate

6.6

7%

7.0

0%

8.0

0%

6.6

3%

7.0

0%

8.0

0%

6.7

2%

7.0

0%

8.0

0%

60

%S

ME

sas

reta

il4

0%

SM

Es

asco

rpo

rate

6.3

1%

6.8

0%

8.0

0%

6.2

3%

6.8

0%

8.0

0%

6.3

0%

6.8

0%

8.0

0%

70

%S

ME

sas

reta

il3

0%

SM

Es

asco

rpo

rate

5.9

5%

6.6

0%

8.0

0%

5.8

4%

6.6

0%

8.0

0%

5.8

8%

6.6

0%

8.0

0%

80

%S

ME

sas

reta

il2

0%

SM

Es

asco

rpo

rate

5.5

9%

6.4

0%

8.0

0%

5.4

4%

6.4

0%

8.0

0%

5.4

6%

6.4

0%

8.0

0%

90

%S

ME

sas

reta

il1

0%

SM

Es

asco

rpo

rate

5.2

4%

6.2

0%

8.0

0%

5.0

5%

6.2

0%

8.0

0%

5.0

4%

6.2

0%

8.0

0%

10

0%

SM

Es

asre

tail

0%

SM

Es

asco

rpo

rate

4.8

8%

6.0

0%

8.0

0%

4.6

5%

6.0

0%

8.0

0%

4.6

2%

6.0

0%

8.0

0%

32 ALTMAN AND SABATO

4. Conclusions

This paper has addressed the subject of the possible effects that the implementation of

the new Basel Capital Accord will have on banks and SMEs. Surveying several other

studies focusing on a similar theme and adding new methodologies, we compare and

expand upon the received evidence and conclusions with several important findings.

First, credit risk models have been developed specifically for SMEs in order to specify

one-year PDs for samples from three different countries in the most recent and relevant

Basel II compliant way. For several of the other inputs needed in the new Basel formulas,

realistic assumptions have been made. Our assumptions, while reasonable, have been

further analyzed by applying a sensitivity analysis to the critical variables.

Second, the new Basel formulas, the ones contained in the last version of the Accord

(June 2004), have been utilized to calculate the requirements. Indeed, for the first time,

new formulas where expected losses (PDxLGD) are subtracted from the capital

requirements have been tested.

Third, our findings confirm, to some extent, what has also been found in the other

studies: i.e., that the part of SMEs classified as retail can enjoy significantly lower capital

requirements than the part classified as corporate. However, our results also show that, if

banking organizations should consider their entire SME portfolio as corporate and

utilizing the A-IRB approach, they will likely face higher capital requirements than

under the Basel I Capital Accord. Banks may be inclined to treat SMEs as corporate,

even for relatively small exposures, since there are additional organizational and

technological requirements to treat them as retail customers. Now the trade-off between

lower capital requirements and higher organizational costs will be clearer and better

evaluated.

The results of our analysis show that, for all three of the countries, a minimum of

20% of small and medium sized enterprises must be classified as retail in order to

maintain the SME capital requirement at least at the current level (8%). Then, we

find that the percentage of SMEs to be considered as retail should be at least 40% if

the banking organizations will want to enjoy a lower capital requirement under the

Basel II Advanced IRB approach versus the Standardized approach. We acknowledge

that banks, developing and implementing their own internal models, will likely find

slightly different breakevens from the one proposed in this work. However, having

developed and applied three models to estimate PDs, we have found a very similar

distribution of SMEs for all of our countries amongst different rating classes. Indeed,

we believe that the relationship between SMEs and credit risk can be considered, on

average, as similar to the ones shown in this work and we expect the range of possible

breakeven values to be distributed in close proximity in either side of our values (20%

and 40%).

We conclude that one of the main results of the new Basel Capital Accord will be to

motivate banks to update their internal systems and procedures in order to be able to

manage SMEs on a pooled basis through the use of a scoring, rating or some other

automated decision system. Only in this way will A-IRB banking organizations receive

major benefits in terms of lower capital requirements and more efficient and profitable

management of the SME portfolio. We argue that especially large, well-diversified

EFFECTS OF THE NEW BASEL CAPITAL ACCORD 33

banks, which will be the first to implement the A-IRB approach, will reap the biggest

benefits, probably reducing (as Berger, 2004, argues) or possibly even eliminating the

current competitive advantage of small, specialized banking organizations based on their

privileged Brelationship lending^ situation. This reduction or possibly even elimination

of the competitive advantage will likely be the case in the US where smaller banks will

almost surely not change from the present Basel I structure. The exact result is difficult

to estimate since it will mainly be a function of customers’ reactions to pricing changes

(if any).

Lastly, for SMEs, themselves, we point out that access to bank financing is likely to

become easier and possibly cheaper, since large banks will find SME lending more

profitable. Through the use of advanced credit risk management techniques an even more

attractive capital allocation will be possible. We conclude that worries about an increase

in the costs of SME finance, mainly due to Basel II implementation costs sustained by

banks, can also be valid, but the use of the most advanced risk-based pricing techniques

should result in higher prices primarily only toward lower quality firms and only during

the initial years after incurring these costs.

Appendix A: Italian SMEs model development

Construction of the data sample

In order to estimate one year PDs for Italian SMEs, the necessary data for the statistical

analysis has been provided by a large Italian bank. The original data set consisted of

about 86,000 firm-year observations spanning the time period 2000 to 2003. These have

been extracted from the loan portfolio of all customers (except private individuals) that

had an exposure less than "1 million and sales less than "50 million. First, we excluded

sole traders, 62,520 clients, since we did not have enough variables to develop a specific

model for them. Then, due to missing data items, the final data set contained 20,193

small and medium sized firms.

To create the binomial dependent variable (default/no default), we observed the

situation of each firm at the end of the next financial year. Following Basel II definition,

only if the company was 90 or more days past due, was it considered as a default. The

distribution of the companies in the data set is shown in table A-1.

Table A-1. Final data set for Italian SMEs

In this table, the structure of the Italian sample used to

develop the model is shown.

Number Percentage

Good firms 19,392 96.5%

Defaults 701 3.5%

Total 20,193 100%

34 ALTMAN AND SABATO

Selection of candidate variables

In the next step, candidate variables for the final model have been selected. Using

univariate analysis, about 40 accounting ratios have been analyzed and 20 of them

selected as the most important and predictive. These have been chosen to be considered

in the logistic regression to develop the final model (table A-2).

Finally, each of the selected ratios was studied from a credit risk point of view, testing

its linearity and its behavior in predicting bankruptcy. All of them were consistent with

our expectations and with the ones found in the literature [See Hayden (2003), for a

complete analysis of the most important accounting ratios and their expected behavior in

predicting bankruptcy. See also Tamari (1966) and Edmister (1972)].

The logit regression

Amongst statistical models, at least five main methods could be used to develop default

prediction models [See Dimitras, Zanakis and Zopoundis (1996), for a literature review

about most of these methods]:

� Linear or non-linear discriminant analysis

� Linear regression

� Logit and probit regression

� Neural networks

� Recursive partitioning analysis.

Considering the quality of the input variables and the opinions expressed by many

studies [see Crouhy et al. (2001)], a logit regression has been chosen.

Using a stepwise variable selection procedure, based on a likelihood-ratio test with

the significance level set at 20%, eight variables were selected in the final model (table

A-3].

These variables can be grouped into three categories describing different aspects of

firms’ structure (table A-4).

Table A-2. List of selected variables

In this table, the variables used in the logistic regression to develop the final model are shown.

Bank debt /(Total assets j Bank debt) Financial profits j Financial charges /Debt

Debt /Equity Cash flow/Short term dank debt

Added value /Total assets Bank debt/Current assets j Current liabilities

Cash /Total assets Equity/Bank debt

Long term liabilities /Total assets Financial charges/Debt

Tangible assets /Total assets ROE

Account Payable /Total assets Financial charges /Liabilities

Long term bank debt /Bank debt Extra-ordinary profits /Ordinary profit

Sales /Current assets Sales growth

Ordinaty profits + Depreciation /debt Assets growth

EFFECTS OF THE NEW BASEL CAPITAL ACCORD 35

Model performance

Figure A-1 shows the cumulative accuracy profile of the developed model. The curved

line shows the performance of the model being evaluated in depicting the percentage of

defaults captured by the model at different percentages of the data set, while the thin

straight line below represents the naıve case of zero information or random assignment of

Table A-3. Variables entered in the final model

In this table, the variables entered in the final model are listed.

Debt/Equity

Bank debt/(Total assets j Bank debt)

Long term liabilities/Total assets

Economic value added/Total assets

Cash/Total assets

Tang. assets/Total assets

Accounts payable/Total assets

Long term bank debt/Bank debt

Table A-4. Categories of the variables

In this table, the variables entered in the final model are grouped on the

basis of the different aspect of the firms’ structure that they describe.

Debt/Equity Leverage

Bank debt/(Total assets j Bank debt)

Long term liabilities/Total assets

Economic value added/Total assets Profitability

Cash/Total assets Liquidity

Tang. assets/Total assets

Accounts payable/Total assets

Long term bank debt/Bank debt

Figure A-1. Cumulative accuracy profile of the Italian model. In this figure, the cumulative accuracy profile

of the Italian model is shown. The curved line shows the performance of the model being evaluated in depicting

the percentage of defaults captured by the model at different percentages of the data set, while the thin straight

line below represents the naıve case of zero information or random assignment of default probabilities.

36 ALTMAN AND SABATO

default probabilities [see Kraft et al. (2004) and Engelman et al. (2003), for further

discussions about how to measure the performance of a model].

Appendix B: US SMEs model development

Construction of the data sample

In order to estimate one-year PDs for US SMEs, the necessary data for the statistical

analysis has been extracted from WRDS Compustat. The original data set consisted of

about 69,000 firm-year observations spanning the time period 2000 to 2003, for firms

with sales less than $65 million. Missing data caused us to drop the majority of firms, but

there were still 12,739 available observations. Then, to develop the model, we selected

only the active and not defaulted companies at the beginning of 2001 and we used their

financial data to predict which of them would have defaulted at the end of the year. The

final sample had 3,048 not defaulted firms and 25 defaulted (table B-1). To create the

binomial dependent variable (default /no default), we observed the footnote field of

Compustat, considering as default firms listed as BTL^ or BAG^.35 Since the low number

of defaults is due only to the bad quality of the data, a weight has been used in the

regression to increase the power of their information.

Selection of candidate variables

We considered the Altman Z-Score (Altman, 1993) model (table B-2), for manufacturing

and non-manufacturing firms, and analyzed the variables within a logistic regression

structure.

The logit regressionVModel performance

Using a statistical stepwise variable selection procedure, based on a likelihood-ratio test

with the significance level set at 20%, after five steps, all the variables were inserted in

the model.

Figure B-1 shows the cumulative accuracy profile of the developed model. The curved

line shows the performance of the model being evaluated in depicting the percentage of

defaults captured by the model at different percentages of the data set, while the thin

straight line below represents the naıve case of zero information or random assignment of

default probabilities.

35 BTL^ footnote is used to indicate firms in bankruptcy, while BAG^ means that the firm is in reorganization.

EFFECTS OF THE NEW BASEL CAPITAL ACCORD 37

Appendix C: Australian SMEs model

For the Australian case, we used the data provided by the Corporate Scorecard Group.

The model applied to these SMEs was the Corporate Scorecard Credit Rating (CSCR)

derived by the ZETA-Score model (Altman et al., 1977). The ZETA-Score model utilizes

the following seven variables:

X1 = EBIT/total assets

X2 = stability of earnings

X3 = EBIT/total interest payments

Figure B-1. Cumulative accuracy profile of the US model. In this figure, the cumulative accuracy profile of

the US model is shown. The curved line shows the performance of the model being evaluated in depicting the

percentages of defaults captured by the model at different percentages of the data set, while the thin straight

line below represents the naıve case of zero information or random assignment of default probabilities.

Table B-1. Final data set for US SMEs

In this table, the structure of the US sample used to develop

the model is shown.

Number Percentage

Good firms 3,048 99.2%

Defaults 25 0.8%

Total 3,073 100%

Table B-2. List of selected variables

In this table, the variables entered in the final model are listed.

Working capital/Total assets

Retained earnings/Total assets

EBIT/Total assets

Book value equity/Book value of total liabilities

38 ALTMAN AND SABATO

X4 = retained earnings/total assets

X5 = current ratio

X6 = equity/total capital

X7 = total assets.

This model is widely accepted in Australia, demonstrating a high degree of default/

non-default accuracy.

Appendix D: Sensitivity analysis

In order to apply the new Basel formulas for the A-IRB approach, we had to provide four

inputs: probability of default (PD), loss given default (LGD), exposure at default (EAD)

and maturity (M). As discussed in the text, we developed and applied specific models to

estimate PD. To further test our assumptions, a sensitivity analysis was performed to test

the effect on our results of using different levels of LGD or M. For LGD, we always use

the same value for the recovery rate for all the rating classes, even if PDs are different.

The values of LGD that we test have a symmetric distance of 20 percentage points from

the average value that we used in the paper (45%).

In tables D-1 and D-2, the significant importance of LGD in the new Basel formulas is

shown. We observe that for low values of LGD, e.g. 25%, all company types (SME

retail, SME corporate and large corporate) demonstrate significant reductions from the

current Basel I requirement (8%) and from the use of a Standardized approach (6%Y8%).

But, for large LGDs (e.g.( 65%), a higher percentage of SMEs (between 70% and 80%)

will be required to be classified as retail in order to enjoy lower capital requirements by

applying the A-IRB instead of the Standardized approach.

Table D-1. Results with LGD = 25%

In this table, a value of 25% for LGD is used to derive bank capital requirements.

Italy US Australia

SMEs as retail 2.71% 2.58% 2.56%

SMEs as corporate 4.70% 4.78% 4.90%

CorporateVLarge firms n.a. 3.69% 4.34%

Table D-2. Results with LGD = 65%

In this table, a value of 65% for LGD is used to derive bank capital requirements.

Italy US Australia

SMEs as retail 7.05% 6.71% 6.67%

SMEs as corporate 12.21% 12.43% 12.73%

CorporateVLarge firms n.a. 9.60% 11.30%

EFFECTS OF THE NEW BASEL CAPITAL ACCORD 39

In tables D-3 and D-4, we show the results changing the value of the maturity (from 3

and 5 years to 1 and 3 years and 5 and 10 years). First, it is clear that the maturity has a

little effect on the capital requirements and it does not affect at all the part of SMEs

considered as retail. However, this analysis also points out that for longer maturities,

banks should increase the part of SMEs classified as retail in order to maintain a lower

level of capital requirements.

References

Advanced Notice of Proposed Rulemaking (ANPR) for the Application of Basel II to U.S. Banking

organizations. (August 2003), http://www.federalreserve.org.

Allen, Linda, and Anthony Saunders. BSpecial Issue on FThe Path to Basel II Adoption._^ Journal of Financial

Services Research 26, no. 2 (2004), 101Y191.

Altman, Edward I. BFinancial Ratios, Discriminant Analysis and the Prediction of Corporate Bankruptcy.^Journal of Finance 23, no. 4 (1968), 589Y611.

Altman, Edward I. Corporate Financial Distress and Bankruptcy. 2nd edn, New York: John Wiley & Sons, 1993.

Altman, Edward I. BCorporate Credit Scoring Insolvency Risk Models in a Benign Credit and Basel II

Environment.^ NYU Working Paper (2004).

Altman, Edward I., and Anthony Saunders. BAn Analysis and Critique of the BIS Proposal on Capital

Adequacy and Ratings.^ Journal of Banking and Finance 25, no. 1 (2001), 25Y46.

Altman, Edward I., R.G. Haldeman, and P. Narayanan. BZeta-analysis. A New Model to Identify Bankruptcy

Risk of Corporations.^ Journal of Banking and Finance 1 (1977), 29Y54.

Altman, Edward I., John Hartzell and Matthew Peck. BA Scoring System for Emerging Market Corporate

Debt.^ Salomon Brothers (May 15, 1995).

Basel Committee on Banking Supervision. BThe Internal Rating-Based Approach.^ (2001), http://www.bis.org.

Basel Committee on Banking Supervision. 2002 BQuantitative Impact Study 3.^ (2002), http://www.bis.org.

Basel Committee on Banking Supervision. BThe New Basel Capital Accord.^ (2003), http://www.bis.org.

Table D-3. Results with maturity of 1 and 3 years

In this table, a one year maturity is used only for small firms (with sales less than "25 million), while a three

year maturity is used for medium (sales between "25 and "50 million) and large firms (sales more than "50

million).

Italy US Australia

SMEs as retail 4.88% 4.65% 4.61%

SMEs as corporate 8.20% 8.32% 8.52%

CorporateVLarge firms n.a 6.12% 7.34%

Table D-4. Results with maturity of 5 and 10 years

In this table, a five year maturity is used only for small firms (with sales less than "25 million), while a ten year

maturity is used for medium (sales between "25 and "50 million) and large firms (sales more than "50 million).

Italy US Australia

SMEs as retail 4.88% 4.65% 4.61%

SMEs as corporate 8.80% 9.05% 9.34%

CorporateVLarge firms n.a 8.49% 9.28%

40 ALTMAN AND SABATO

Basel Committee on Banking Supervision. BInternational Convergence of Capital Measurement and Capital

Standards.^ (2004), http://www.bis.org.

Berger, Allen N. BPotential Competitive Effects of Basel II on Banks in SME Credit Markets in the United

States.^ (2004), http://www.federalreserve.org.

Berger, Allen N., and Gregory F. Udell. BCollateral, Loan Quality, and Bank Risk.^ Journal of Monetary

Economics 25 (1990), 21Y42.

Berger, Allen N., and Gregory F. Udell. BRelationship Lending and Lines of Credit in Small Firm Finance.^Journal of Business 68 (1995), 351Y382.

Berger, Allen N., and Gregory F. Udell. BThe Economics of Small Business Finance: The Roles of Private Equi-

ty and Debt Markets in the Financial Growth Cycle.^ Journal of Banking and Finance 22 (1998), 613Y673.

Berger, Allen N. and Gregory F. Udell. BA More Complete Conceptual Framework about SME Finance.^World Bank Conference on BSME: Overcoming Growth Constraints,^ October 14Y15, 2004.

Bocchi, Lorenzo and Giuseppe Lusignani. BLe Nuove Regole di Basilea II: Prime Valutazioni di Impatto sul

Rapporto Banca-Impresa in Italia.^ Banca Impresa Societa, 23, no. 2 (2004).

Booth, J.R. BContract Costs, Bank Loans and the Cross-Monitoring Hypothesis.^ Journal of Financial

Economics 31 (1992), 25Y41.

Crouhy, M., D. Galai, and R. Mark. BA Comparative Analysis of Current Credit Risk Models.^ Journal of

Banking and Finance 24 (2001), 59Y117.

Dimitras, A., S. Zanakis, and C. Zopoundis. BA Survey of Business Failures with Emphasis on Prediction

Methods and Industrial Applications.^ European Journal of Operational Research 90 (1996), 487Y513.

Edmister, R. BAn Empirical Test of Financial Ratio Analysis for Small Business Failure Prediction.^ Journal of

Financial and Quantitative Analysis (1972), 1477Y1493.

Engelman, Berndt, Evalyn Hayen, and Dirk Tasche. BMeasuring the Discriminative Power of Rating Systems.^(2003), http://www.defaultrisk.com.

European Commission, Commission Recommendation 96/280/EC of April 3, 1996. http://europa.eu.int.

European Commission, Commission Recommendation 2003/361/EC of May 6, 2003. http://europa.eu.int.

European Commission, Enterprise Publications. B2003 Observatory of European SME.^ http://europa.eu.int.

Gibson, Brian. BDefinition of Small Business.^ University of Newcastle (2001).

Grunert, Jens, Lars Norden, and Martin Weber. BThe Role of Non-Financial Factors in Internal Credit Ratings.^Journal of Banking and Finance 29 (2004), 509Y531.

Hannan, Timothy H., and Steven J. Pilloff. BWill the Adoption of Basel II Encourage Increased Bank Merger

Activity? Evidence from the United States.^ (2004), http://www.federalreserve.org.

Hayden, Evelyn. BAre Credit Scoring Models Sensitive with Respect to Default Definition? Evidence from the

Austrian Market.^ (2003), http://www.defaultrisk.com.

Heitfield, Erick. BRating System Dynamics and Bank-Reported Default Probabilities under the New Basel

Capital Accord.^ (2004), http://www.defaultrisk.com.

Kolari, James W., and Hwan G. Shin. BAssessing the Profitability and Riskiness of Small Business Lenders in

the U.S. Banking Industry.^ (2004), http://www.sba.gov.

Kraft, Holger, Gerald Kroisandt, and Marlene Muller. BRedesigning Ratings: Assessing the Discriminatory

Power of Credit Scores under Censoring.^ (2004), http://www.defaultrisk.com.

Jacobson, Tor, Jesper Linde, and Kasper Roszbach. BCredit Risk Versus Capital Requirements under Basel II:

Are SME loans and Retail Really Different?^ Journal of Financial Services Research 28, no. 1/2 (2005),

43Y75.

Lopez, J.A. BThe Empirical Relationship between Average Asset Correlation, Firm Probability of Default and

Asset Size.^ Journal of Financial Intermediation 13, no. 2 (2004), 265Y283.

PricewaterhouseCoopers and NIESR Research for European Commission. BStudy on the Financial and

Macroeconomic Consequences of the Draft Proposed New Capital Requirements for Banks and Investment

Firms in the UE.^ (2004), http://europa.eu.int.

Repullo, Rafael, and Javier Suarez. BLoan Pricing Under Basel Capital Requirements.^ Centre for Economic

Policy Research Discussion Paper 3917 (March 2004).

Saurina, Jesus and Carlos Trucharte. BThe Impact of Basel II on Lending to Small-and-Medium-Sized Firms: A

Regulatory Policy Assessment Based on Spanish Credit Register Data.^ Journal of Finance Services

Research 26 (2004), 121Y144.

EFFECTS OF THE NEW BASEL CAPITAL ACCORD 41

Schwaiger, Walter S.A. BBasel II: Quantitative Impact Study on Austrian Small and Medium-sized

Enterprises.^ Technical University of Vienna (2002).

Tamari, M. BFinancial Ratios as a Means of Forecasting Bankruptcy.^ Management International Review

(1966), 15Y21.

The Challenge, World Bank Review of Small Business Activities. (2001), http://www.ifc.org.

Udell, Gregory F. BSME Lending: Defining the Issues in a Global Prospective.^ Indiana University Working

Paper (2004).

42 ALTMAN AND SABATO